Antag Weighting Feedback

[Haprenti]

Suggestion, the weighting system uses a rolling 30 day window. This would help with veterans getting abysmally low chances while also avoiding having to re-prime the system from something like a monthly clean slate

The system would need to also take into account how much of the past month you've actually as well, otherwise new people would still have an advantage

A problem I forsee though is antag fishing

It could also, instead of being based on the number of times you've played as "such" antag, be based on the percentage of rounds you've played as that antag.

For instance, if you want player to ideally have 2% of their rounds as a wraith (Taking into account ideally set % of rounds played as an antag, and % of antag rounds played as a wraith, I'm taking that number out of my ass obviously)

Let's say you've played for a long time and managed to have 30 rounds as a wraith, but that amount to 1% of your rounds.
You'd still have more chance to play as a wraith compared to a guy that only played once as a wraith, but has played only one round.
That other guy had 100% of his rounds playing as a wraith, and an incredibly small weight for this selection (Not taking into account the 10% chance to ignore those of course.).

Edit : Meant 2% instead of 1% in the example, fixed

[Wire]

That's a good idea! It will require a small refactor in terms of what player data I'm tracking (essentially I'll need more data) but very do-able.

[Wire]

I have done the percentage suggestion thing!

Now instead of being based purely on lowest plays, it will select players with the lowest percentage of rounds played as the given antag type. So, for example, if player A has 2% of their traitor-rounds played as traitor, they will get chosen above someone who has 10%.

Note that it tracks % picked for a given round mode (traitor, wizard, etc) so it should be pretty darn fair (e.g. no skewing the probabilities if a dude plays a lot of blob mode exclusively).

This should be reasonable hopefully maybe?? Let me know!

[OMJ]

I swear I've been rolling antag a LOT more since this has been implemented

[Ed Venture]

I swear I've been rolling antag a LOT more since this has been implemented

I got Traitor once last week so it feels as normal as ever for me.

[Wire]

Like I said it needs a decent amount of data to really "feel" different. Also I had to wipe the records for this percentage refactor so we're starting from scratch again. Wooo.

[awfulworldkid]

Does this take into account rounds when I had Traitor turned off?

[warcrimes]

Does this take into account rounds when I had Traitor turned off?

if it does, people could effectively "bank" chances when they just wanna have chill researchy rounds, without missing out on potential future antag stuff. sounds feature.

[Wire]

Yep that's how it would work ("banking" chances) which I'm going to declare a feature because of the reason you gave and also it's less work to leave it.

[Superlagg]

I really like the idea of banking antag chances. Sounds like a fair way to make it more likely for people who actually *want* to be antags to be antags without sprinkling tokens all over.

[warcrimes]

Yeah my thoughts exactly. It follows in line with the whole process of turning certain roles off- someone who doesnt want to be a certain antag is gonna be a shitty antag- and someone who's spent a dozen rounds learning some new stuff and wants to try to put it to use, that's gonna be a pretty interesting traitor. Sounds like a win-win. Can't wait to see how this pans out.

[Sundance]

Banking antagonist rounds cumulatively over rounds is a great approach.

[Rhydic]

it would be nice if we could view our % (maybe even vs server average over some period of time) but i can see how that might be a problem for any of a number of reasons

[Wire]

Yeah I want to avoid any chance of gaming the system so I think it would be best to keep those numbers a secret.

[warcrimes]

So far it feels a lot better, but that might just be confirmation bias.